1

DOI 10.1007/s11284-014-1152-3 1

Link to published journal article: 2

https://link.springer.com/article/10.1007%2Fs11284-014-1152-3 3

4

5

Synergistic effects of primates and dung beetles on soil seed accumulation in snow 6

regions 7

8

Hiroto ENARI1,*, Haruka SAKAMAKI-ENARI2 9

10

1Faculty of Agriculture, Yamagata University, 1-23 Wakabamachi, Tsuruoka, Yamagata 11

997-8555, Japan 12

2Faculty of Agriculture, Iwate University, 3-18-3 Ueda, Morioka, Iwate 020-8550, Japan 13

14

*Communicating author: E-mail: enari@tds1.tr.yamagata-u.ac.jp; Tel & Fax: 15

+81-235-28-2925 16

17

2

Abstract 18

This study aimed to reveal the soil seed accumulation processes for endozoochorous 19

plants in the heavy-snowfall forests of Japan, where seed dispersal agents are few when 20

compared to tropical forests. We assessed (1) primary seed dispersal by Japanese 21

macaques (Macaca fuscata) by identifying dispersed seeds found in their feces, and (2) 22

secondary seed dispersal by dung beetles by using beads (as seeds mimics) of different 23

sizes, to quantify the frequency of seed burial and burial depths. We studied this 24

diplochorous system in different forest types (undisturbed beech forest, conifer 25

plantation, and secondary beech-oak forest) and during the spring and summer seasons. 26

The key findings were as follows: (1) macaques dispersed the seeds of 11 and 14 plant 27

species during spring and summer, respectively; (2) seeds dispersed by macaques in the 28

spring were smaller and twice as abundant than those dispersed in the summer; (3) 29

although no differences were observed in the amount of beads buried by beetles 30

between seasons, all bead sizes tended to be buried in deeper soil layers in the spring 31

than in the summer; and (4) the seed supply to the soil in undisturbed beech forest and 32

conifer plantation was greater than the one in secondary beech-oak forest. Similar to 33

what has been observed in tropical forests, seeds defecated by frugivorous mammals 34

can be successfully incorporated into the underground soil seed bank through a 35

diplochorous macaque-beetle system in temperate forests of deep snow regions. 36

37

Key words Biological interaction · Conifer · Fagus crenata · Macaca fuscata · 38

Scarabaeinae Seed dispersal 39

40

3

Introduction 41

When investigating the demography of a plant, we tend to focus only on the shoot 42

system, the plant body that is aboveground, because of the constraints of our visibility. 43

However, some plants possess a much larger population as seeds in the soil seed bank 44

compared to the population of stems in an area (Cook 1980; Fenner 1985; Silvertown 45

and Charlesworth 2001). This difference becomes more pronounced for early 46

successional or ruderal plants, which usually have longer seed longevity than late 47

successional or mature forest species (Bossuyt and Hermy 2001; Decocq et al. 2004). 48

The abundance and species composition of soil seeds has a marked influence not only 49

on the dynamics of plant populations and communities, but also on the resilience of 50

ecosystems to natural catastrophes (Bakker et al. 1996; Hopfensperger 2007). 51

Formation of a soil seed bank for endozoochorous plant species requires primary 52

seed dispersal by vertebrates in order to escape the area of high seed mortality near the 53

parent plant, and may be further facilitated through secondary seed dispersal by other 54

animals that transport seeds to sites that are safe from seed predators (Vander Wall and 55

Longland 2004, 2005; Dalling 2005; Feer et al. 2013). Rodents and granivorous ants are 56

known to act as secondary seed dispersers because of their scatter-hoarding behavior, 57

but they are also recognized as seed predators (Forget 1996; Retana et al. 2004, Vander 58

Wall et al. 2005; Koike et al. 2012a). Other ant types do not eat the seeds, but they 59

collect only elaiosome-bearing seeds, in particular those that have been dispersed by 60

birds (Böhning-Gaese et al. 1999; Vander Wall et al. 2005). Dung beetles (Coleoptera: 61

Scrabaeinae), on the other hand, have been identified as being superior secondary seed 62

dispersers for mammal-defecated seeds. Dung beetles transport many seeds 63

4

underground where they are safe from seed predation; these insects are commonly rich 64

in biomass among decomposer organisms throughout temperate and tropical regions 65

(Andresen and Feer 2005; Nichols et al. 2008). 66

Dung beetle diversity and biomass have been shown to be highly sensitive to 67

changes in mammal fauna (resource [feces] providers) both in tropical biomes (Hanski 68

and Cambefort 1991; Nichols et al. 2009), and in temperate forests of heavy snowfall 69

regions (Enari et al. 2013). The latter type of ecosystem supports mammals in limited 70

abundance and species richness, resulting in corresponding limited biomass and 71

diversity of beetles when compared to those in warmer forests (Hanski and Cambefort 72

1991; Davis et al. 2002; Enari et al. 2011, 2013). These facts logically indicate that the 73

agents contributing to multistep seed dispersal systems, such as diplochory (Vander 74

Wall and Longland 2004, 2005) may be limited in regions that experience heavy snow, 75

which might lead to low functional redundancy for sustaining seed dynamic processes 76

for plant species dispersed by frugivorous mammals. 77

With the exception of a few studies conducted in temperate zones—grassland 78

(D’hondt et al. 2008) and forests with light snowfall (Koike et al. 2012b)—all of the 79

studies focusing on the ecology and behavior of dung beetles as part of a diplochorous 80

system have been conducted in the tropics (Andresen and Feer 2005; Nichols et al. 81

2008). As a result, we have little knowledge of the ecological processes involved in the 82

formation of a soil seed bank for plants with seeds dispersed trough endozoochory in 83

ecosystems with deep snow and limited agents of seed dispersal. This study intended to 84

evaluate the diplochorous system consisting of primary seed dispersal by Japanese 85

macaques (Macaca fuscata) and secondary seed dispersal by dung beetles in the 86

5

Shirakami Mountains, which are located in one of the heaviest snowfall regions in the 87

world. The macaque is a primate species adapted for survival in cold and heavy snow 88

conditions (Enari 2014) and is regarded as both a key seed disperser (Otani 2010; Tsuji 89

et al. 2011) and resource provider for dung beetles (Enari et al. 2011, 2013). Compared 90

to fruiting phenology in the tropics, that in temperate forests drastically varies with the 91

the seasons, resulting in rapid fluctuations in fruit availability for animals (Hanya et al. 92

2013). Moreover, the pronounced changes in weather conditions among seasons results 93

in restriction of the active period of most dung beetles, especially in deep snow regions 94

where they appear mainly between June and September (Enari et al. 2011, 2013). The 95

exact duration of their activity period also depends strongly on the type of forest cover 96

(Enari et al. 2011). Thus, in this study we discuss the ecological features of the 97

diplochorous system in terms of the influences of seasonality and forest type, as well as 98

its net effect on the local ecosystem process. 99

Methods 100

Study area 101

We conducted the current study in the northeastern part of the Shirakami Mountains, 102

at the northern end of mainland Japan (central coordinates = 40°32′9″N, 140°14′43″E). 103

The study area has a cool-temperate climate with heavy snowfall from mid-November 104

to late March, with snow cover reaching depths of 2 m in the low-lying flatlands and >4 105

m in montane regions. Lingering snow is usually observed until mid-May, meaning that 106

the soil surface is covered by snow for half of the year. The southwestern side of the 107

study area falls within the Shirakami Mountains World Heritage Site (designated by 108

UNESCO), where undisturbed forests of beech (Fagus crenata) occur. In contrast, 109

6

anthropogenically disturbed broadleaf forests of beech and oak (Quercus crispula), 110

because of fuel wood productions, and conifer plantations (Japanese cedar, Cryptomeria 111

japonica) are irregularly distributed in the northeastern side of the study area. According 112

to the weather station at the Shirakami-sanchi World Heritage Conservation Center, the 113

mean air temperature and the total precipitation are 18.4 ± 1.7°C (SD) 92.8 ± 21.4 mm 114

in June, and 23.6 ± 3.8°C and 185.6 ± 67.5 mm in August, respectively (c.f., annual 115

mean, 10.0 ± 1.3°C and 1,679.8 ± 262.3 mm). 116

Japanese macaques are distributed throughout the study area. In 2008, individual 117

and troop densities in and around the study area were 5.2 individuals km-2 and 0.2 troop 118

km-2, respectively (Enari and Sakamaki 2011). The area is inhabited by five 119

medium-sized or large frugivorous/omnivorous mammals other than macaques: Martes 120

melampus, Mustela itatsi, Nyctereutes procyonoides, Meles anakuma, and Ursus 121

thibetanus. Among these, U. thibetanus is widely known as an effective seed disperser 122

in terms of dispersal quantity and dispersal distance (Koike et al. 2008, 2011), but in 123

deep snow regions its feces are less attractive to dung beetles compared to macaque 124

feces (Enari et al. 2013). 125

Of the 152 dung beetle species that have been recorded for the Japanese 126

archipelago (Kawai et al. 2005) the authors have confirmed the presence of 15 species 127

in the Shirakami Mountains (Enari et al. 2011, 2013). Dung beetles are usually divided 128

into three functional groups according to nesting strategy and dung-relocation behavior: 129

dwellers, tunnellers, and rollers (Cambefort and Hanski 1991). Only the latter two 130

groups bury feces in tunnels underground (Estrada and Coates-Estrada 1991; Feer 1999). 131

In the Shirakami Mountains six tunneller species that use macaque feces are present 132

7

(Onthophagus ater, O. nitidus, O. fodiens, O. atripennis, Phelotrupes auratus, P. 133

laevistriatus), but no rollers that use macaque feces exist (Enari et al. 2011, 2013). The 134

abundance of most species is notably high in spring regardless of forest type (Enari et al. 135

2011). 136

Primary seed dispersal by macaques 137

To collect fresh feces we captured three adult female macaques from different 138

troops in the study area using box traps. All individuals were fitted with radio collars 139

(Advanced Telemetry Systems, Isanti, Minnesota, USA; weight = 120 g, approximately 140

1% of the average of the three females’ body mass) and released. All manipulations 141

were in accordance with the Wildlife Protection and Hunting Law of Japan. We 142

followed the movement of each troop and collected fresh feces during spring (June–July 143

2007 and June–July 2008) and summer (August–September 2007). We washed each 144

sample with running water through a 0.5-mm mesh sieve to extract the seeds and 145

identified each seed species following Nakayama et al. (2006). 146

Secondary seed dispersal by dung beetles 147

During spring and summer of 2012 we evaluated secondary dispersal by dung 148

beetles of seeds in macaque feces, in three forest types where the seasonal compositions 149

of existing dung beetle assemblages had been previously investigated (Enari et al. 150

2011): undisturbed mature beech forests with a closed canopy (age of ca. 140 y), conifer 151

plantations with a closed canopy (ca. 40 y) and secondary beech-oak forests with an 152

open canopy (ca. 50 y). The soil type at all sites was brown forest soil. We measured 153

soil hardness, which may influence the ability of beetles to bury feces, 60 times at each 154

site using a Yamanaka soil hardness tester (Fujiwara Scientific Co., Ltd., Tokyo, Japan). 155

8

All measurements were taken on the same day. We found soil hardness values of 1.99 ± 156

3.37 (SD) kg cm-2 in undisturbed beech forests, 1.29 ± 2.19 kg cm-2 in conifer 157

plantations, and 1.60 ± 0.85 kg cm-2 in secondary beech-oak forests—there was no 158

significant difference among sites (One-way ANOVA, F2, 177 = 1.29, P = 0.28). 159

To assess seed burial by dung beetles we used fresh macaque feces collected by 160

following the troops, as mentioned previously. After removing dung beetles found on 161

the feces, we homogeneously mixed all of the collected feces and prepared 15-g fecal 162

piles, the mean weight of macaque feces in the Shirakami Mountains (Enari et al. 2011). 163

We then placed ellipsoidal plastic beads, used as seed mimics, of three sizes into each 164

fecal pile: 10 large beads (length [L] = 6.1 mm, thickness [T] = 5.9 mm), 15 medium 165

beads (L = 4.1 mm, T = 3.9 mm), and 30 small beads (L = 2.2 mm, T = 1.4 mm). The 166

size and abundance of beads used were chosen according to what is naturally found in 167

macaque feces in cool-temperate forests (Tsuji et al. 2011). Next, we set the 168

bead-containing fecal piles (with a total of 55 beads each) along a 40-m transect at 10-m 169

intervals (i.e., five fecal piles). One transect was used in each site, on two dates: June 22, 170

2012 and August 16, 2012. To protect the experimental fecal piles from natural 171

disturbances (such as rain and rodents), we covered each pile with a wire cage (30 cm × 172

20 cm × 20 cm; 1-cm mesh size) with a roof. The mesh size used allowed free passage 173

of all dung beetles observed in cool-temperate forests of Japan (Koike et al. 2012a). 174

After one month we collected soil samples within a radius of 10 cm from each fecal pile. 175

Samples were collected at 5 cm increments at depths between 0 and 30 cm, i.e., from 176

Layer 0 (0 – 5 cm depth) to Layer 5 (25 – 30 cm). The maximum depth was determined 177

according to Koike et al. (2012b), who reported that dung beetles can burrow to a depth 178

9

of up to 29 cm in temperate forests. Some beads might have been buried simply by 179

falling into natural soil crevices. Therefore, we considered only beads found in Layers 1 180

– 5 as having been buried through dung beetle activity. We recovered beads from every 181

soil sample using 1-, 3-, and 5-mm mesh sieves and compared the mean percentages of 182

beads buried in the different soil layers at each forest type and between seasons. 183

Data analyses 184

To identify areal and seasonal differences of primary seed dispersal by macaques, 185

we compared the frequency of macaque defecations found in the different forest types 186

by using Bonferroni z statistic (Neu et al. 1974) and the weight of macaque feces and 187

the abundance of seeds in the feces in spring vs. summer by using Mann–Whitney U 188

test. We also evaluated the abundance of seeds of different sizes in spring vs. summer 189

by using Welch’s t test. We then quantified secondary seed dispersal by dung beetles on 190

the basis of the frequency of seed burial and burial depth. To analyze the frequency of 191

seed burial, we compared the percentage of beads buried by beetles (i.e., beads found in 192

layers 1 to 5) in the three forest types by using a significance test for multiple 193

comparisons of proportions (Ryan 1960). To reveal seasonal trends in burial depths of 194

beads with respect to bead size, we compared the abundance of beads buried in different 195

soil layers by using the Steel–Dwass multiple comparison test (Steel 1960). To assess 196

the validity of these multiple comparisons under the current sampling design, we 197

calculated the achieved statistical power (i.e., 1– error probability) for each 198

significance test by using the post-hoc power analysis with effect size f ( = 0.05; 199

Cohen 1988). We performed all these procedures by using R ver. 3.0.2 (R Core 200

Development team). 201

10

Results 202

Seed dispersal by macaques 203

We collected 88 fecal samples in spring, of which 18.2% were from beech forest; 204

22.7%, from conifer plantation; 34.1%, from secondary broadleaf forest; and the 205

remaining mainly from riverbeds. We collected 81 fecal samples in summer, of which 206

24.7% were from undisturbed beech forest; 16.0%, from conifer plantation; 30.9%, 207

from secondary broadleaf forest; and the remaining mostly from the edges of farmlands. 208

The frequency of macaque defecations in the different forest types showed no 209

significant difference during both spring (Bonferroni z statistic: n.s., 2 = 4.7, z = 2.1) 210

and summer (n.s., 2 = 3.7, z = 2.1). The mean weight of macaque feces was similar in 211

both seasons (spring: 12.8 ± 12.5 g, N = 88; summer: 12.4 ± 8.1 g, N = 81; 212

Mann–Whitney U test: z = -1.42, P = 0.15). While we did not observe a seasonal 213

difference in the percentages of seed occurrence in the feces (86.3% in the spring and 214

86.4% in the summer), mean seed abundance per feces was significantly different: 215

106.0 ± 215.9 seeds in the spring vs. 53.8 ± 102.4 in the summer (Mann–Whitney U 216

test: z = 2.25, P = 0.02). 217

In the feces collected during the spring, we found seeds of 11 plant species 218

(Appendix 1). Among the observed species (excluding those that we were unable to 219

identify to the species level) were two tall trees, five shrubs, and one liana. In the feces 220

collected during the summer we found 14 seed species. Excluding the three species that 221

were identified only up to family level, we found seeds from two tall trees, three shrubs, 222

and six lianas. 223

While most seeds found during spring were smaller than the size of the medium 224

11

beads (i.e., 4.1 mm), seeds found in the summer varied more in size (Table 1; Appendix 225

1). All seed species found in macaque feces in both seasons, with the exception of two 226

species, were shade-tolerant or early successional species (Forestry Development 227

Technological Institute 1985; Satake et al. 1993; Appendix 1) whose seeds are 228

commonly found in persistent soil seed banks (Nakagoshi 1985; Thompson 1992) 229

surviving >1 year (Mizui 1993). 230

Secondary seed dispersal by dung beetles 231

The disappearance of the fecal piles during the study period was 100% for every 232

experimental fecal pile, both in the spring and summer seasons. The percentages of 233

beads that we were able to recover were 90.7 ± 10.6% in the spring and 94.7 ± 8.1% in 234

the summer for large beads, 81.8 ± 15.3% in the spring and 85.8 ± 13.7% in the summer 235

for medium beads, and 72.9 ± 18.7% in the spring and 86.7 ± 11.0% in the summer for 236

small beads. 237

In total, dung beetles buried 28.5% and 39.4% of the experimental beads (based on 238

recovered beads) in soil layer >5-cm deep during spring and summer, respectively. 239

However, the percentage of beads buried by beetles showed different trends among 240

forest types depending on season (Table 2). In spring, bead burial was significantly 241

higher in undisturbed beech forests, regardless of bead size (Ryan’s multiple 242

comparisons test; P < 0.05), followed by conifer plantations. In summer, bead burial in 243

undisturbed beech forests were not always the highest, while burial in conifer 244

plantations were constantly high for every bead size (Table 2). According to the 245

post-hoc power analysis, the value for every achieved statistical power was more than 246

0.8, which supports the validity of the current sampling design for these comparisons 247

12

(Cohen 1988). 248

The mean percentage of beads buried in the different soil layers varied seasonally 249

(Fig. 1). This evaluation led to the following findings: (1) regardless of forest type and 250

season, larger beads were rarely buried in deeper soil layers; (2) although dung beetles 251

tended to disperse beads in relatively shallow soil layers (mostly Layer 1) during the 252

summer, they could bury beads in deeper soil layers during the spring; and (3) the trend 253

of (2) became more obvious for medium and small beads in both undisturbed beech 254

forests and conifer plantations, which was supported by the Steel–Dwass multiple 255

comparison test. When showing the results of the power analysis, most values of the 256

achieved statistical power for large beads was less than 0.50, meaning a deficiency in 257

sample size or low effect size, which was inferred from the fact that most large beads 258

were found in Layer 0 (i.e., surface soil). 259

Discussion 260

Soil seed accumulation process 261

The current findings show that in our study region, a region with heavy snow and 262

limited seed dispersal agents, seeds defecated by macaques are highly likely to be 263

incorporated into the underground soil seed bank through secondary dispersal by dung 264

beetles. In fact, the percentage of beads buried by dung beetles in our study is similar to 265

values reported in several studies conducted in the tropics (range = 25–52%; see 266

Chapman 1989; Andresen and Feer 2005 and references therein). Further, the 267

percentage of seed burial by beetles we estimated is likely to be an underestimate of the 268

real value, given our conservative inclusion criterion (i.e., only beads found at depths >5 269

cm were treated as beads buried by dung beetles). On the other hand, it should also be 270

13

noted that our methods removed the influence of granivorous animals, and thus we are 271

presenting seed burial by dung beetles in the absence of seed removal by other biotic 272

factors. Seed removal by granivorous animals, in particular rodents, however, might 273

have a limited influence on seed burial by beetles in our study site, because large beetles 274

(Phelotrupes spp.), which have been shown to have a superior ability for burying seeds 275

in the temperate forests of Japan (Koike et al. 2012b), have the same daily rhythm as 276

macaques (i.e., diurnal; Enari et al. 2011), whereas most rodents are nocturnal. This 277

indicates that Phelotrupes spp. probably find and bury feces before the risk of seed 278

predation by rodents increases (Koike et al. 2012b). 279

Studies have repeatedly reported that the percentage of smaller seeds buried by 280

beetles is higher than that of larger seeds (Estrada and Coates-Estrada 1991; Shepherd 281

and Chapman 1998; Feer 1999; Andresen 1999, 2002; Andresen and Levey 2004; 282

Andresen and Feer 2005; Pouvelle et al. 2009; Feer et al. 2013). Moreover, several 283

studies conducted in the tropics demonstrated that seed burial depth is also inversely 284

correlated with seed size (Shepherd and Chapman 1998; Andresen 2002; Pouvelle et al. 285

2009; Feer et al. 2013). The current findings are consistent with those previous works. 286

Considering that the seeds dispersed by macaques in the spring are smaller but twice as 287

numerous as seeds dispersed in the summer (Table 1), the present findings suggest that 288

seed supply to the soil through the plant-macaque-beetle interaction might be higher in 289

the spring than in the summer. 290

Regarding the vertical soil layer distribution of buried seeds, there seems to be a 291

difference between tropical forests and the deep snow forests of Japan. While previous 292

studies in the tropics have shown that most seeds buried by dung beetles are found at 293

14

depths < 10 cm (Shepherd and Chapman 1998; Andresen 2001, 2002; Andresen and 294

Levey 2004), our study results show that beetles in the spring frequently buried seeds in 295

deeper soil layers (>10 cm) regardless of seed size (Fig. 1). Yet, this comparison should 296

be made with care, considering that seed sizes included in tropical studies tend to be 297

larger, on average, than those used in our study. 298

Besides the relatively small seed sizes used in our study, the observed burial depths 299

might also be partly explained by the seasonal abundance of beetles in this region, 300

where the timing of the occurrence of most beetles overlaps during a brief period, i.e., 301

June (Enari et al. 2011, 2013). We also note that this seasonal prevalence is more 302

obvious for dweller species (Enari et al. 2011). Dwellers do not directly bury feces 303

underground but may change the shape and consistency of the feces, which might 304

promote dung and seed burial by tunnellers. 305

Although there was no remarkable difference in the percentage of bead burial 306

between spring and summer (Table 2), all bead sizes tended to be buried more shallowly 307

during the summer (Fig. 1). This, again, may be explainable in terms of the seasonal 308

abundance of different dung beetles species, in particular tunneller species, the size of 309

which is positively correlated with the size of the seeds that they can bury underground 310

(Feer 1999; Andresen 2002). In our study region, every tunnellers emerge in spring. 311

Among them, small tunnellers (Onthophagus spp.), which can bury seeds up to depths 312

of 10 cm (Koike et al. 2012b), are active in summer, while large tunnellers (Phelotrupes 313

spp.) with deep burial ability of up to 30 cm (Koike et al. 2012b) are less common in 314

summer, except in conifer plantations with closed canopy (Enari et al. 2011; Enari H 315

2014, unpublished data). 316

15

The consequences of burial for seed fate naturally vary with burial depth. Seeds 317

buried in deeper soil layers are least vulnerable to predation by rodents (Janzen 1982; 318

Estrada and Coates-Estrada 1991; Andresen 1999; Andresen and Levey 2004). Similary 319

one experiment conducted in temperate forests reported that seeds buried deeper than 2 320

cm are rarely detected by rodents (Koike et al. 2012b). Thus, it may be reasonable to 321

suppose that the burial activity of beetles in the summer still generates sufficient 322

benefits for seeds in terms of the avoidance of predation risk. On the other hand, burial 323

depth also influences seed longevity. Seed longevity in storage generally lengthens with 324

lower temperatures and moisture contents (Roberts 1972; Murdoch and Ellis 1992); 325

therefore, seeds buried deeper in the ground commonly have longer longevity (Toole 326

and Brown 1946; Dalling et al. 1998), although this physiological trait may vary with 327

species and habitat conditions. Unfortunately, there have been few empirical studies to 328

quantify the relationships between seed burial depth and longevity in temperate forests, 329

so it remains to be tested whether the storage environment of seeds buried in the 330

summer is suitable for long-term survival. 331

The current study did not directly focus on seed dispersal in autumn because it is an 332

inactive season for dung beetles. Although macaques still act as primary seed dispersers 333

during autumn (Otani 2010; Tsuji et al. 2011), dispersed seeds are probably under a 334

higher risk of predation. Such predation risk, however, might be diminished by the 335

presence of snowfall in late autumn. Given that macaque feces containing seeds of 336

autumnal fruits are preserved under snow cover until the next spring, those autumnal 337

feces possibly are buried underground by spring-emerging beetles (Enari H and 338

Sakamaki-Enari H 2013, unpublished data). 339

16

Forest type differences in soil seed input 340

The composition of dung beetle assemblages is highly sensitive to forest type and 341

habitat disturbance in the tropics (Vulinec 2000; Andresen 2003; Andresen and Feer 342

2005) and also in heavy snow regions (Enari et al. 2011). Our results suggest that soil 343

seed input varies with forest patches of different landscapes when dung beetle 344

assemblages change (Enari et al. 2011), even if those patches are geographically close 345

(Fig. 1; Table 2). Such local variations in seed supply could be driven by macaque 346

habitat use or its feces supply, which determines not only the primary seed input but 347

also the biomass of dung beetle serving as seed buriers. In fact, similar primate–beetle 348

chain effects have also been confirmed in the tropics (Vulinec et al. 2006; Pouvelle et al. 349

2009; Feer et al. 2013). 350

Not only the current vegetation, but also historical land use generally determines 351

the constituents of soil seeds in a subject area (Bossuyt and Hermy 2001). However, 352

such legacy effects tend to fade within 50 years after the formation of an existing stable 353

landscape because of the longevity of buried seeds (Bossuyt and Hermy 2001). This 354

means that the richness of soil seeds in climax forests observed in the heavy snow 355

regions—i.e., beech forests with few woody plants producing persistent soil 356

seeds—becomes depleted without a new seed supply (excluding beech) from outside the 357

forest patch. Our results (Fig. 1; Table 2) suggest that fresh seeds transported by 358

macaques (Appendix 1) constantly accumulate as soil seeds in climax forests. 359

Conifer plantations have often been dismissed in terms of biodiversity conservation 360

because of the monotone landscape they create (Hunter 1990). Moreover, considering 361

that approximately 40% of conifer plantations in Japan are >50 years old (according to 362

17

the public database from the Forestry Agency of Japan), a strong legacy effect on soil 363

seed accumulations can no longer be expected. The current experiments regarding 364

primary and secondary seed dispersals demonstrate that macaques provide stable soil 365

seed inputs even in conifer plantations, especially during the summer (Table 2; Fig. 1). 366

Unlike other large mammals in the region, macaques repeatedly use such plantations as 367

an alternative habitat, especially for their sleeping sites, throughout the year (Sakamaki 368

and Enari 2012; Enari and Sakamaki-Enari 2013, 2014). A study conducted in the 369

tropics shows a similar cascading phenomenon, referred to as the latrine effect 370

(Pouvelle et al. 2009). This study showed that forest areas used as sleeping sites by 371

primates (i.e., defecation areas) possess abundant seeds and high seed diversity (in 372

particular, small seeds belonging to pioneer tree species) in the soil. Unfortunately, the 373

details of the latrine effect are yet to be investigated in cool-temperate forests. However, 374

this effect might be a key ecosystem function of the processing-chain commensalism 375

beginning with macaques by supporting future generations of diverse organisms in 376

plantations. 377

Acknowledgements 378

We wish to thank Issei Nakahara, Takafumi Kanno, and Kana Okuda-Nomoto; 379

citizen volunteers recruited by Earthwatch Japan; and rangers at the Shirakami-sanchi 380

World Heritage Conservation Center for assisting us with fieldwork and experiments. 381

All research reported in this paper strictly adhered to the current laws of Japan. This 382

work was supported by funds from the Japan Society for the Promotion of Science 383

grants-in-aid programs for fellows (19–5374) and scientific research (23710279) and by 384

the Mitsui & Co., Ltd., Environment Fund. 385

18

References 386

Andresen E (1999) Seed dispersal by monkeys and the fate of dispersed seeds in a 387

Peruvian rain forest. Biotropica 31:145-158. doi: 10.2307/2663968 388

Andresen E (2001) Effects of dung presence, dung amount and secondary dispersal by 389

dung beetles on the fate of Micropholis guyanensis(Sapotaceae) seeds in Central 390

Amazonia. J Trop Ecol 17:61-78. doi: 10.1017/S0266467401001043 391

Andresen E (2002) Dung beetles in a Central Amazonian rainforest and their ecological 392

role as secondary seed dispersers. Ecol Entomol 27:257-270. doi: 393

10.1046/j.1365-2311.2002.00408.x 394

Andresen E (2003) Effect of forest fragmentation on dung beetle communities and 395

functional consequences for plant regeneration. Ecography 26:87-97. doi: 396

10.1034/j.1600-0587.2003.03362.x 397

Andresen E, Feer F (2005) The role of dung beetles as secondary seed dispersers and 398

their effect on plant regeneration in tropical rainforests. In: Forget PM, Lambert JE, 399

Hulme P, Vander Wall S (eds) Seed fate: predation, dispersal and seedling 400

establishment. CABI Publishing, Oxon, UK, pp. 331-349 401

Andresen E, Levey D (2004) Effects of dung and seed size on secondary dispersal, seed 402

predation, and seedling establishment of rain forest trees. Oecologia 139:45-54. doi: 403

10.1007/s00442-003-1480-4 404

Böhning-Gaese K, Gaese BH, Rabemanantsoa SB (1999) Importance of primary and 405

secondary seed dispersal in the Malagasy tree Commiphora guillaumini. Ecology 406

80:821-832. doi: 10.1890/0012-9658(1999)080[0821:IOPASS]2.0.CO;2 407

Bakker J, Poschlod P, Strykstra R, Bekker R, Thompson K (1996) Seed banks and seed 408

19

dispersal: important topics in restoration ecology. Acta Bot Neerl 45:461-490 409

Bossuyt B, Hermy M (2001) Influence of land use history on seed banks in European 410

temperate forest ecosystems: a review. Ecography 24:225-238. doi: 411

10.1034/j.1600-0587.2001.240213.x 412

Cambefort Y, Hanski I (1991) Dung beetle population biology. In: Hanski I, Cambefort 413

Y (eds) Dung beetle ecology. Princeton University Press, Princeton, pp 36-50 414

Chapman CA (1989) Primate seed dispersal: the fate of dispersed seeds. Biotropica 415

21:148-154. doi: 10.2307/2388705 416

Cook R (1980) The biology of seeds in the soil. In: Solbrig OT (ed) Demography and 417

evolution in plant populations. University of California Press, California, pp 418

107-130 419

Dalling JW, Swaine M, Garwood NC (1998) Dispersal patterns and seed bank dynamics 420

of pioneer trees in moist tropical forest. Ecology 79: 564-578. doi: 421

10.1890/0012-9658(1998)079[0564:DPASBD]2.0.CO;2 422

Dalling JW (2005) The fate of seed banks: factors influencing seed survival for 423

light-demanding species in moist tropical forests. In: Forget PM, Lambert JE, 424

Hulme P, Vander Wall S (eds) Seed fate: predation, dispersal and seedling 425

establishment. CABI Publishing, Oxon, UK, pp. 31-44 426

Davis ALV, Scholtz CH, Philips TK (2002) Historical biogeography of scarabaeine 427

dung beetles. J Biogeogr 29:1217-1256. doi: 10.1046/j.1365-2699.2002.00776.x 428

Decocq G, Valentin B, Toussaint B, Hendoux F, Saguez R, Bardat J (2004) Soil seed 429

bank composition and diversity in a managed temperate deciduous forest. Biodivers 430

Conserv 13:2485-2509. doi: 10.1023/B:Bioc.0000048454.08438.C6 431

20

D’hondt B, Bossuyt B, Hoffmann M, Bonte D (2008) Dung beetles as secondary seed 432

dispersers in a temperate grassland. Basic Appl Ecol 9: 542-549. doi: 433

10.1016/j.baae.2007.11.002 434

Enari H (2014) Snow tolerance of Japanese macaques inhabiting high-latitude 435

mountainous forests of Japan. In: Grow N, Gursky-Doyen S, Krztn A (eds) High 436

altitude primates. Springer, New York, pp 133-151 437

Enari H, Koike S, Sakamaki H (2011) Assessing the diversity of dung beetle 438

assemblages utilizing Japanese monkey feces in cool-temperate forests. J For Res 439

16:456-464. doi: 10.1007/s10310-010-0236-3 440

Enari H, Koike S, Sakamaki H (2013) Influences of different large mammalian fauna on 441

dung beetle diversity in beech forests. J Insect Sci 13:54. doi: 442

10.1673/031.013.5401 443

Enari H, Sakamaki-Enari H (2013) Resource use of Japanese macaques in heavy 444

snowfall areas: implications for habitat management. Primates 54:259-269. doi: 445

10.1007/s10329-013-0343-9 446

Enari H, Sakamaki-Enari H (2014) Impact assessment of dam construction and forest 447

management for Japanese macaque habitats in snowy areas. Am J Primatol 448

76:271-280. doi: 10.1002/ajp.22231 449

Enari H, Sakamaki H (2011) Estimation of abundance and distribution of Japanese 450

macaques using track-counts in snow. Acta Theriologica 56:255-265. doi: 451

10.1007/s13364-011-0025-1 452

Estrada A, Coates-Estrada R (1991) Howler monkeys (Alouatta palliata), dung beetles 453

(Scarabaeidae) and seed dispersal: ecological interactions in the tropical rain forest 454

21

of Los Tuxtlas, Mexico. J Trop Ecol 7:459-474. doi: 10.1017/S026646740000585X 455

Feer F (1999) Effects of dung beetles(Scarabaeidae) on seeds dispersed by howler 456

monkeys(Alouatta seniculus) in the French Guianan rain forest. J Trop Ecol 457

15:129-142. doi: 10.1017/S0266467499000711 458

Feer F, Ponge JF, Jouard S, Gomez D (2013) Monkey and dung beetle activities 459

influence soil seed bank structure. Ecol Res 28:93-102. doi: 460

10.1007/s11284-012-1006-9 461

Fenner M (1985) Seed ecology. Chapman & Hall, London 462

Forestry Development Technological Institute (1985) The knowledge of useful 463

broadleaf trees. Taiheisya, Tokyo 464

Forget P-M (1996) Removal of seeds of Carapa procera (Meliaceae) by rodents and 465

their fate in rainforest in French Guiana. J Trop Ecol 12:751-761. doi: 466

10.1017/S0266467400009998 467

Hanski I, Cambefort Y (1991) Species richness. In: Hanski I, Cambefort Y (eds) Dung 468

beetle ecology. Princeton University Press, Princeton, pp 350-365 469

Hanya G, Tsuji Y, Grueter CC (2013) Fruiting and flushing phenology in Asian tropical 470

and temperate forests: implications for primate ecology. Primates 54: 101-110. doi: 471

10.1007/s10329-012-0341-3 472

Hopfensperger KN (2007) A review of similarity between seed bank and standing 473

vegetation across ecosystems. Oikos 116:1438-1448. doi: 474

10.1111/j.2007.0030-1299.15818.x 475

Hulme P, Kollmann J (2005) Seed predator guilds, spatial variation in post-dispersal 476

seed predation and potential effects on plant demography: a temperate perspective. 477

22

In: Forget PM, Lambert JE, Hulme P, Vander Wall S (eds) Seed fate: predation, 478

dispersal and seedling establishment. CABI Publishing, Oxon, UK, pp. 9-30 479

Hunter ML (1990) Wildlife, forests, and forestry: principles of managing forests for 480

biological diversity. Prentice Hall, Englewood Cliffs, New Jersey 481

Janzen DH (1982) Removal of Seeds from Horse Dung by Tropical Rodents: Influnce 482

of Habitat and Amount of Dung. Ecology:1887-1900. doi: 10.2307/1940128 483

Kawai S, Hori S, Kawahara M, Inagaki M (2005) Atlas of Japanese Scarabaeoidae, 484

vol.1 Coprophagous group. Roppon-Ashi Entomological Books, Tokyo 485

Koike S et al. (2011) Estimate of the seed shadow created by the Asiatic black bear 486

Ursus thibetanus and its characteristics as a seed disperser in Japanese 487

cool-temperate forest. Oikos 120:280-290. doi: 10.1111/j.1600-0706.2010.18626.x 488

Koike S, Morimoto H, Goto Y, Kozakai C, Yamazaki K (2008) Frugivory of carnivores 489

and seed dispersal of fleshy fruits in cool-temperate deciduous forests. J For Res 490

13:215-222. doi: 10.1007/s10310-008-0069-5 491

Koike S et al. (2012a) Seed removal and survival in Asiatic black bear Ursus thibetanus 492

faeces: effect of rodents as secondary seed dispersers. Wildl Biol 18:24-34. doi: 493

10.2981/10-049 494

Koike S et al. (2012b) The role of dung beetles as a secondary seed disperser after 495

dispersal by frugivore mammals in a temperate deciduous forest. Acta Oecol 496

41:74-81. doi: 10.1016/j.actao.2012.04.009 497

Mizui N (1993) Ecological studies on reproduction in deciduous broad-leaved tree 498

species. Bull Hokkaido Forestry Res Inst 30:1-67 (In Japanese with English 499

abstract) 500

23

Murdoch AJ, Ellis RH (1992) Longevity, viability and dormancy. In: Fenner M (ed) 501

Seeds: the ecology of regeneration in plant communities. CABI, Oxon, UK, pp 502

193-229 503

Nakagoshi N (1985) Buried viable seeds in temperate forests. In: White J (ed) The 504

population structure of vegetation. Junk, Dordrecht, pp 551-570 505

Nakayama S, Inokuchi M, Minamitani T (2006) Seeds of wild plants in Japan-second 506

edition. Tohoku University Press, Sendai 507

Neu CW, Byers CR, Peek JM (1974) Technique for analysis of utilization-availability 508

data. J Wildl Manage 38:541-545. 509

Nichols E, Gardner T, Peres C, Spector S (2009) Co-declining mammals and dung 510

beetles: an impending ecological cascade. Oikos 118:481-487. doi: 511

10.1111/j.1600-0706.2009.17268.x 512

Nichols E, Spector S, Louzada J, Larsen T, Amezquita S, Favila M (2008) Ecological 513

functions and ecosystem services provided by Scarabaeinae dung beetles. Biol 514

Conserv 141:1461-1474. doi: 10.1016/j.biocon.2008.04.011 515

Otani T (2010) Seed dispersal by Japanese macaques. In: Nakagawa N, Nakamichi M, 516

Sugiura H (eds) The Japanese macaques. Springer, Tokyo, pp 129-142 517

Retana J, Xavier Pico F, Rodrigo A (2004) Dual role of harvesting ants as seed predators 518

and dispersers of a non-myrmechorous Mediterranean perennial herb. Oikos, 519

105:377-385. doi: 10.1111/j.0030-1299.2004.12854.x 520

Pouvelle S, Jouard S, Feer F, Tully T, Ponge JF (2009) The latrine effect: impact of 521

howler monkeys on the distribution of small seeds in a tropical rain-forest soil. J 522

Trop Ecol 25: 239-248. doi: 10.1017/S0266467409005987 523

24

Roberts EH (1972) Storage environment and the control of viability. In: Roberts EH 524

(ed) Viability of seeds. Chapman and Hall, London, pp 14-58 525

Ryan TH (1960) Significance tests for multiple comparison of proportions, variances, 526

and other statistics. Psychological Bulletin 57: 318-328. 527

Sakamaki H, Enari H (2012) Activity-specific evaluation of winter habitat use by 528

Japanese macaques in snow areas, northern Japan: implications for conifer 529

plantation management. For Ecol Manage 270:19-24. doi: 530

10.1016/j.foreco.2012.01.010 531

Satake Y, Hara H, Watari S, Tominari T (1993) Wild flower of Japan. Heibonsya, Tokyo 532

Shepherd V, Chapman C (1998) Dung beetles as secondary seed dispersers: impact on 533

seed predation and germination. J Trop Ecol 14:199-215. doi: 534

10.1017/S0266467498000169 535

Silvertown JS, Charlesworth D (2001) Introduction to plant population biology-4th. ed. 536

Blackwell Science, Oxford 537

Steel, RGD (1960) A rank sum test for comparing all pairs of treatments. Technometrics 538

2:197-207. doi:10.1080/00401706.1960.10489894 539

Thompson, K (1992) The functional ecology of seed banks. In: Fenner M (ed) Seeds: 540

the ecology of regeneration in plant communities. CABI, Oxon, UK, pp. 235-258 541

Toole EH, Brown E (1946) Final results of the Duvel buried seed experiment. J Agric 542

Res 72:201-210 543

Tsuji Y, Sato K, Sato Y (2011) The role of Japanese macaques (Macaca fuscata) as 544

endozoochorous seed dispersers on Kinkazan Island, northern Japan. Mammalian 545

Biol 76:525-533. doi: 10.1016/j.mambio.2011.01.001 546

25

Vander Wall SB, Kuhn KM, Beck MJ (2005) Seed removal, seed predation, and 547

secondary dispersal. Ecology 86:801-806. doi: 10.1890/04-0847 548

Vander Wall SB, Longland WS (2004) Diplochory: are two seed dispersers better than 549

one? Trends Ecol Evol 19:155-161. doi:10.1016/j.tree.2003.12.004 550

Vander Wall SB, Longland WS (2005) Diplochory and the evolution of seed dispersal. 551

In: Forget PM, Lambert JE, Hulme P, Vander Wall S (eds) Seed fate: predation, 552

dispersal and seedling establishment. CABI, Oxon, UK, pp. 297-314 553

Vulinec K (2000) Dung beetles (Coleoptera: Scarabaeidae), monkeys, and conservation 554

in Amazonia. Fla Entomol:229-241 555

Vulinec K, Lambert JE, Mellow DJ (2006) Primate and dung beetle communities in 556

secondary growth rain forests: Implications for conservation of seed dispersal 557

systems. Int J Primatol 27:855-879. doi: 10.1007/s10764-006-9027-2 558

Figure legends

Fig. 1 Seasonal variation in the percentage (mean ± SE) of bead buried in soil layers of

different depths: Layer 1, 5-10 cm; Layer 2, 10-15 cm; Layer 3, 15-20 cm; Layer 4, 20-25cm;

Layer 5, 25-30 cm. Asterisks above the bars show significant differences among layers

according to the Steel-Dwass test, P < 0.05. The values between parentheses are achieved

statistical power (i.e., 1 – error probability), calculated by the post-hoc power analysis ( =

0.05), and effect size f

Table 1 Differences in the length of seeds observed in Japanese macaque feces in the Shirakami Mountains, Japan

Seeds Large (> 4.1 mm)a Small (< 4.1 mm) Welch’s t test

Spring (N = 88) Abundance (/feces)b 25.5 ± 33.1 80.3 ± 219.3 t = −2.32, P = 0.02

Number of species 3 > 8c NA

Summer (N = 81) Abundance (/feces)b 19.5 ± 26.9 34.0 ± 103.7 t = −1.22, P = 0.22

Number of species > 7 > 7 NA

aMean length of seeds according to Nakayama et al. (2006). Threshold value was determined based on the length of medium beads

bMean ± standard deviation (SD)

cThe symbol “>” denotes the class containing unidentified plant seeds

Table 2 Influence of forest type on bead burial (> 5 cm burial depth) by dung beetles during spring and summer, for large, medium and

small beads used as seed mimics (see text for bead sizes). Forest types are: undisturbed beech forest (BF), conifer plantation (CP) and

secondary beach-oak forest (SF).

Large beads Medium beads Small beads

BF CP SF BF CP SF BF CP SF

Spring

% beads burieda 37.2+b 13.3 10.4 55.8+b 21.9 17.6 47.8+b 25.0 25.0

(1– c; Effect size f) (0.86; 0.99) (1.00; 1.52) (0.99; 1.37)

Summer

% beads burieda 11.1 36.2+b 4.0 61.0 45.2 19.4–b 56.0 60.8 27.8–b

(1– c; Effect size f) (1.00; 1.91) (1.00; 1.82) (1.00; 5.42)

a The percentages of beads buried were calculated after excluding unrecovered beads

b + and – signs represent significantly higher and lower, respectively, compared to the other two forest types, according to Ryan’s

multiple comparisons test (P < 0.05)

c Achieved statistical power, calculated by the post-hoc power analysis ( = 0.05)

Undisturbedbeechforests(0.99; 1.18)

Coniferplantations

(0.97; 1.02)

Secondarybeech-oakforests(0.94; 0.95)

Undisturbedbeechforests(1.00; 3.32)

Coniferplantations

(1.00; 2.78)

Secondarybeech-oakforests(1.00; 1.28)

0

25

50

0

25

50

0

25

50

Layer 1Layer 2Layer 3Layer 4Layer 5

(a) Spring, large beads (b) Summer, large beads

(c) Spring, medium beads (d) Summer, medium beads

(e) Spring, small beads (f) Summer, small beads

Forest type

Perc

enta

ge o

f bea

ds b

urie

d

**

** **

* **

(0.66; 0.67) (0.60; 0.64) (0.27; 0.41) (1.00; 1.94) (1.00; 1.40) (0.63; 0.65)

(0.48; 0.56) (0.24; 0.39) (0.10; 0.22) (0.12; 0.25) (0.98; 1.07) (0.06; 0.10)